Interestana
Home/News/Amazon Facility Destroys Rare Books for AI Training
Ars Technica2 min read

By Interestana AI Editorial — AI-drafted, human-overseen. How we report

Amazon Facility Destroys Rare Books for AI Training

Amazon Facility Destroys Rare Books for AI Training

Amazon has been revealed to be destroying rare books, including those with significant historical or cultural value, as part of its efforts to train artificial intelligence models. This practice came to light when a bookseller, suspecting AI firms were acquiring and then destroying books after scanning them, agreed to plant an Apple AirTag in a rare book that was part of a bulk order. The AirTag was subsequently tracked to an Amazon AI training facility located in Las Vegas. According to reports by 404 Media, this facility housed a dedicated team whose task involved removing books from their spines and scanning individual pages. The discovery provides concrete evidence supporting long-standing suspicions within the bookselling community regarding the methods employed by some technology giants in their pursuit of advanced AI capabilities. The race to develop frontier AI models has apparently led to the destruction of valuable literary works, raising significant ethical and preservation concerns.

The facility's operations were further underscored by a logo displayed on the door of the team's warehouse, identified as VGT3. This logo depicted a Tyrannosaurus rex in the act of devouring a book, a visual metaphor that critics found to be particularly insensitive given the escalating backlash against destructive book scanning practices. This imagery suggests a deliberate and perhaps even defiant approach to the process of data acquisition for AI training. While the exact nature of the AI models being trained and the specific datasets being compiled remain undisclosed by Amazon, the physical destruction of rare books points to a potentially unsustainable and ethically questionable method of data procurement. The implications extend beyond the immediate loss of these artifacts, potentially impacting the availability of historical texts for future research and public access.

Amazon has largely deflected direct comment on these findings. When approached by media outlets, including Ars Technica and 404 Media, the company provided a standardized statement that did not specifically address the AI training aspect or the destruction of books. This statement was the same one provided to 404 Media, which initially broke the story. The lack of a detailed response from Amazon leaves many questions unanswered regarding the scale of this operation, the criteria for selecting books for destruction, and the company's long-term strategy for AI data acquisition. The incident highlights a critical tension between technological advancement and the preservation of cultural heritage, prompting calls for greater transparency and accountability from AI developers concerning their data sourcing practices. The use of a hidden AirTag to expose this practice demonstrates the lengths to which individuals and organizations are going to uncover and report on potentially harmful industry practices.

Original source — read the full reporting at the publisher:

Read on Ars Technica

Get the weekly AI digest

AI news + new model releases, weekly. Drafted by our agents, reviewed by humans.

Read next